|Home | About | Journals | Submit | Contact Us | Français|
Background: Krüppel-like factor 4 (KLF4) is a zinc-finger transcription factor with diverse regulatory functions in proliferation, differentiation, and development. KLF4 also plays a role in inflammation, tumorigenesis, and reprogramming of somatic cells to induced pluripotent stem (iPS) cells. To gain insight into the mechanisms by which KLF4 regulates these processes, we conducted DNA microarray analyses to identify differentially expressed genes in mouse embryonic fibroblasts (MEFs) wild type and null for Klf4. Methods: Expression profiles of fibroblasts isolated from mouse embryos wild type or null for the Klf4 alleles were examined by DNA microarrays. Differentially expressed genes were subjected to the Database for Annotation, Visualization and Integrated Discovery (DAVID). The microarray data were also interrogated with the Ingenuity Pathway Analysis (IPA) and Gene Set Enrichment Analysis (GSEA) for pathway identification. Results obtained from the microarray analysis were confirmed by Western blotting for select genes with biological relevance to determine the correlation between mRNA and protein levels. Results: One hundred and sixty three up-regulated and 88 down-regulated genes were identified that demonstrated a fold-change of at least 1.5 and a P-value < 0.05 in Klf4-null MEFs compared to wild type MEFs. Many of the up-regulated genes in Klf4-null MEFs encode proto-oncogenes, growth factors, extracellular matrix, and cell cycle activators. In contrast, genes encoding tumor suppressors and those involved in JAK-STAT signaling pathways are down-regulated in Klf4-null MEFs. IPA and GSEA also identified various pathways that are regulated by KLF4. Lastly, Western blotting of select target genes confirmed the changes revealed by microarray data. Conclusions: These data are not only consistent with previous functional studies of KLF4's role in tumor suppression and somatic cell reprogramming, but also revealed novel target genes that mediate KLF4's functions.
Krüppel-like factor 4 (KLF4) [1, 2] is a member of the KLF family of zinc finger transcription factors that are involved in diverse biological processes including proliferation, apoptosis, differentiation, and development [3-7]. KLF4 also plays an important role in pathological conditions such as tumorigenesis and inflammation [8-14]. Moreover, recent studies indicate that KLF4 is involved in the reprogramming of somatic cells to induced pluripotent stem (iPS) cells [15-20]. The finding that KLF4 overexpression prevents mouse embryonic stem (ES) cell differentiation suggests that KLF4 contributes to ES cell self-renewal .
Mice deficient for Klf4 have been generated. Klf4-null (Klf4-/-) mice die shortly after birth and exhibit defects in terminal differentiation of epithelial tissues such as the epidermis and colon [22, 23]. Mice with tissue-specific deletion of Klf4 also have perturbed homeostasis in tissues from which the gene was deleted including the conjunctiva and stomach [24, 25]. In contrast, mice heterozygous for Klf4 (Klf4+/-) are normal but have increased tumor burden in the intestine when bred to ApcMin mice that are genetically predisposed to develop intestinal adenomas . Conversely, inhibition of oncogenic Notch signaling in ApcMin mice results in an increase in Klf4 expression accompanied by a reduction in intestinal tumor burden . These results are highly suggestive of a tumor suppressive function for KLF4 in the intestinal epithelium. Recent studies demonstrating that mouse embryonic fibroblasts (MEFs) null for the Klf4 alleles are genetically unstable as evidenced by the presence of aneuploidy, chromosome aberration, and centrosome amplification are consistent with this notion .
Despite growing evidence that KLF4 mediates many important physiological processes as exemplified above, the biochemical mechanisms by which KLF4 exerts many of its functions are not well established. Previous studies involving transcriptional profiling of KLF4 when it is overexpressed in a colon cancer cell line indicate that KLF4 has a global inhibitory effect on macromolecular biosynthesis and the cell cycle [27, 28]. However, no systemic evaluation has been conducted to examine the global expression profiles of KLF4 in untransformed cells. Here we compared the expression profiles of KLF4 between MEFs wild type and null for the Klf4 alleles in an attempt to gain further insight into the mechanism of action of KLF4 in a physiological context.
Mice heterozygous for the Klf4 alleles (Klf4+/-) on a C57BL/6 background  were crossbred. MEFs wild type (Klf4+/+), heterozygous (Klf4+/-), or null (Klf4-/-) for Klf4 were derived from day 13.5 embryos using the 3T3 protocol as previously described . Briefly, 106 MEFs were plated on 10-cm dishes and maintained in Dulbecco's modified Eagle's Medium (DMEM), supplemented with 10% fetal bovine serum (FBS), and 1% penicillin-streptomycin at 37°C in atmosphere containing 5% CO2. Cells were passed every 3 days at a density of 106 cells per 10-cm dish. The breeding of mice and isolation of MEFs from mice were approved by the Emory University Institutional Animal Care and Use Committee (protocol number 098-2007).
RNA was processed from cells that had reached 80-90 confluency. Total RNA from cultured wild type and Klf4-null MEFs in triplicate was extracted using Trizol reagent as recommended by the manufacturer (Invitrogen; Carlsbad, CA). RNA was subjected to DNase I treatment in order to remove any contaminating genomic DNA. Final purification was performed on RNAeasy columns (Qiagen; Valencia, CA), according to the manufacturer's recommendations. The integrity of total RNA was confirmed by formaldehyde agarose gel electrophoresis. The RNA was quantified by spectrophotometric reading at 260 and 280 nm and RNA with OD260 /OD280 > 1.8 was submitted for microarray analysis.
Purified RNA was shipped to the Emory Bio-marker Service Center, Emory University, Atlanta, GA, for microarray analysis. Concentration of the RNA was quantified by a Nanodrop spec-trophotometer (Wilmington, DE) and quality was assessed using the Agilent Bioanalyzer (Foster City, CA). Samples with the RNA integrity number of > 7 were used for further microarray analysis. RNA was amplified into cRNA and labeled by in vitro transcription using Illumina TotalPrep RNA Amplification Kit (Ambion, Applied Biosystems; Foster City, CA). Samples were then hybridized to the Mouse WG-6 v2.0 Expression Beadchip that queries 45,281 transcripts that cover over 19,000 unique, curated genes in the NCBI RefSeq database (Build 36, Release 22). The chips were processed as per manufacturer's instructions without any modification. The arrays were scanned using the BeadStation 500 Instrument (Illumina Inc.; San Diego, CA) and data were normalized using the GenomeStudio v1.0.2 (Illumina Inc.; San Diego, CA). The data discussed in this publication have been deposited in the National Center for Biotechnology Information (NCBI's) Gene Expression Omnibus (GEO) and are accessible through GEO series accession number GSE21768.
The background subtraction, expression summary, normalization, and log base 2 transformation of gene signals were carried out using Illumina Beadchip software (Illumina Inc.; San Diego, CA). Significant genes were identified using the significance analysis of microarrays (SAM) software , for which 1,000 random class assignment permutations estimated a false discovery rate (FDR) rate of 1%. This resulted in the identification of 6,218 genes with significant changes in expression between Klf4+/+ and Klf4-/- MEFs.The 6,218 differentially expressed genes were annotated and biological processes were analyzed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) (www.david.abcc.ncifcrf.gov). A fold-change of > 1.5 or < -1.5 and P < 0.05 were used as the criteria for significant gene expression changes between the Klf4+/+ and Klf4-/- cells. This narrowed the number of significant genes down to 251 genes, including 163 up-regulated and 88 down-regulated ones, in Klf4-null cells.
Pathway analyses were conducted on the 6,218 differentially expressed genes with a FDR of 1% identified above using Gene Set Enrichment Analysis (GSEA; www.broad.mit.edu/gsea) and Ingenuity Pathway Analysis (IPA; www.ingenity.com). GSEA, based on the Kolmogorov-Smirnov statistic, was performed as described . GSEA is a knowledge-based approach for interpreting genome-wide expression profiles, using 1,000 trials with randomly permuted class label to estimate a P-value. For each gene set, the ES (enrichment score) were normalized to account for differences in gene set size. The false discovery rate (FDR) was then calculated relative to the normalized enrichment score (NES) values to determine the false-positive rate. Significant FDR and P-values were less than 25% and 0.001, respectively, in accordance with GSEA recommendations.
IPA assigns biological functions to genes using the Ingenuity Pathways Knowledge Base (Ingenuity Systems, Inc., Redwood City, CA). In this, genes could be sorted several times to different groups, if their function is known as to be multimodal. The dataset containing the gene identifiers and fold-changes were uploaded into the web-based application and each gene identifier was mapped to its corresponding gene object in the Ingenuity Pathways Knowledge Base. After the analysis, generated biological function genes are ordered by P-value of significance and maximum number of genes.
Following protein extraction, Western blot analysis was conducted using primary antibodies against CDK2, MMP3, SUMO3, and β-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA), STAT3, pSTAT3 and SOCS3 (Cell Signaling, Danvers, MA, USA). The blots were incubated with appropriate horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. The antibody-antigen complex was visualized by ECL chemiluminescence (Amersham, Pittsburgh, PA, USA).
To identify differentially expressed genes between wild type and Klf4-null MEFs, complimentary RNAs in triplicate were hybridized to the Illumina Mouse WG-6 v2.0 Expression BeadChip containing 45,218 probes that represent over 19,000 unique, curated mouse genes in the NCBI RefSeq database (Build 36, Release 22). Significance analysis of microarray (SAM) was used to analyze the original normalized dataset. This revealed a total of 6,218 genes that were differentially expressed in the Klf4-null cells compared to wild type MEFs with a false discovery rate (FDR) equal to or less than 1% (1% chance of genes falsely identified as differentially expressed). Among this group, 163 up-regulated and 88 down-regulated genes in Klf4-null compared to wild type MEFs exhibited at least a 1.5 fold-change in expression levels and a P-value < 0.05. Both the up- and down-regulated differentially expressed genes were submitted to DAVID (Database for Annotation, Visualization and Integrated Discovery), a web-based application (david.abcc.ncifcrf.gov) that allows access to a relational database of functional annotation [32, 33]. Shown in Tables 1 and and22 are examples of the up-regulated and down-regulated genes in Klf4-null cells, respectively, that have identifiable molecular functions. Moreover, many of these genes can be clustered into major functional categories. For example, up-regulated genes in Klf4-null cells encode cell cycle activators, extracellular matrix proteins, proto-oncogenes, growth factors, and proteins involved in ubiquitination and inflammatory responses (Table 1). In contrast, a distinct group of genes is down-regulated in Klf4-null cells and includes those encoding JAK-STAT signaling proteins, homeobox proteins, glutathione metabolism, and ephrins (Table 2). The full lists of up- and down-regulated genes in Klf4 -null MEFs are provided as supplementary materials (Tables S1 and andS2,S2, respectively).
We next used Ingenuity Pathway Analysis (IPA; www.ingenuity.com) to test for enrichment of known gene function. IPA groups significant genes according to biological processes in which they function. The program displays the genes' significance values, the other genes with which it interacts, and how the genes' products directly and indirectly act on each other. The criteria applied for the search of major biological function categories were maximum number of genes and the P-value of significance. A range of P-values between 3.18 ×10-14 to 5.75 × 10-03 is considered statistically significant. Table 3 shows the most significant results of analysis of 6,218 differentially expressed genes with a FDR less than or equal to 1% identified by SAM. As shown, top biological functions regulated by KLF4 include tumorigenesis, cell death, neoplasia, cancer, apoptosis, proliferation and growth of cells. This result is consistent with previous findings that KLF4 is involved in tumor suppression, cellular proliferation, and apoptosis. Interestingly, one particularly large gene set is involved in neurological disorder (1,016 genes) although the P-value just reached statistical significance (Supplementary Table S3).
We also used the GSEA functional enrichment analysis to interrogate molecular pathways enriched in the two MEFs. In this exercise, 6,218 differentially expressed genes were analyzed for gene sets enriched in Klf4-null and wild type cells. A total of 47 pathway gene sets, 23 in Klf4 -null and 24 in wild type cells, were significantly enriched with a P-value < 0.05 and FDR < 0.25. Tables 4 and and55 show the lists of gene sets enriched in Klf4-null and wild type MEFs, respectively. The complete dataset is provided as Supplementary Table S4, which also includes the specific genes in the pathway gene sets that were enriched in the Klf4-null or wild type MEFs. In addition, snapshots of the enriched pathways in Klf4-null and wild type cells are provided as Supplementary Figures S1 and andS2,S2, respectively.
In Klf4-null MEFs, GSEA showed a significant enrichment in gene sets associated with bile acid biosynthesis, hematopoietic cell lineage, multiple myeloma, genes expressed in medulloblastomas, cell motility, cytokine receptor, cell surface receptors, autophagy, and inflammatory response. Moreover, signaling pathways including ERBB, toll like receptors, and hedgehog signaling were enriched in Klf4-null cells. On the other hand, significant enrichment in gene sets associated with antigen processing and presentation, IL-2 receptor pathway, HOX genes, JAK-STAT signaling pathways, glutathione metabolism, basal transcription factors, and adipocytes differentiation were enriched in the wild type cells. Figure 1 shows the results of enrichment of the JAK-STAT signaling pathway in wild type MEFs. In this example, GSEA mapped 134 out of 1,381 genes and found a highly significant correlation between the gene list and dataset (P < 0.001 and FDR q-value = 0.192). This is consistent with the results in Table 2 showing that some of the genes involved in the JAK-STAT signaling pathway such as JAK3, STAT3, and SOCS3, are down-regulated in Klf4-null MEFs. Among these factors, STAT3 is required for embryonic stem cell maintenance and SOCS3 is involved in differentiation of embryonic stem cells [21, 34, 35].
We validated some of the microarray data by Western blot analysis of select genes in wild type and Klf4-null MEFs. In the microarray analysis CDK2, MMP3, and SUMO3 mRNAs were found to be up-regulated in the Klf4-null cells (Table 1). On the other hand, STAT3 and SOCS3 mRNAs were down-regulated (Table 2). Consistent with the microarray observations, Western blot analysis of Klf4-null MEFs showed excellent correlation in changes of expression for each of these genes between wild type and Klf4-null cells (Figure 2). Interestingly, the level of phosphorylated STAT3 (pSTAT3) was also reduced in Klf4-null cells.
Since it was initially identified some 14 years ago [1, 2], KLF4 has been shown to play an increasingly broad and important function in both physiological and pathological processes. Physiologically, KLF4 regulates proliferation, differentiation, development, apoptosis, and somatic cell reprogramming. KLF4 is also involved in disease conditions such as tumorigenesis and inflammation. Earlier studies indicate that KLF4 is a potent inhibitor of cell proliferation [1, 36] and mediates the cell cycle-checkpoint function of the tumor suppressor, p53 [37-40]. Subsequent studies confirmed this inhibitory effect by the demonstration that KLF4 exerts a tumor suppressive effect in vivo [9, 10]. Previous attempts at establishing the expression profiles of KLF4 were conducted in cultured cancer cells over-expressing KLF4 [27, 28]. These studies confirmed the cell cycle-checkpoint activity of KLF4 and provided additional evidence that KLF4 regulates both epithelial differentiation and macromolecular biosynthesis. In contrast, the present study shows for the first time the transcriptional profiles of KLF4 in a non-transformed cell system and as a result, identified many additional novel targets of KLF4 such as those involved in extracellular matrix, ubiquitin, growth factors, chemotaxis, JAK-STAT, and ephrin signaling (Tables 1 and and2).2). Moreover, the current study does not involve over-expression as in the previous work, thus rendering the results more physiologically relevant.
The cells used in the current study, mouse embryonic fibroblasts (MEFs), have previously been characterized . Relative to wild type cells, MEFs deficient for Klf4 had both a higher rate of proliferation and apoptosis. In addition, Klf4-null cells exhibited evidence of genetic instability as evidenced by the presence of aneuploidy, chromosome aberration and centrosome amplification . A mechanism underlying this genetic instability in the absence of KLF4 is likely due to elevated cyclin E levels, which are normally suppressed by KLF4 . The current study provides additional supporting evidence by showing that CDK2 is significantly up-regulated in Klf4-null cells (Table 1 and Figure 2). Accompanying the increase in CDK2 levels is the up-regulation of numerous other cell cycle-promoting genes as shown in Table 1. Moreover, at least a subset of these genes such as MCM and E2F overlaps with those previously identified to be suppressed by KLF4 [27, 28]. The results of the present study therefore provide further mechanistic evidence for the observed inhibitory effect of KLF4 on proliferation.
Consistent with the tumor suppressive role for KLF4, results in Table 1 also show that numerous genes involved in tumorigenesis such as proto-oncogenes and those encoding extracellular matrix proteins, growth factors, chemokines, and inflammatory response. Many of these gene families have important roles in regulating cell growth, migration, and angiogenesis. For example, several genes encoding matrix metalloproteinases (MMPs) are up-regulated in Klf4-null cells (Table 1). The up-regulation of MMPs have has been implicated in the increase in proliferation, anchorage-independent growth, tumor progression, invasion, and metastasis [41-43]. One in particular, MMP3, promotes cellular proliferation when over-expressed in transgenic mice . Over-expression of MMP3 in vitro induces mesenchymal-epithelial transition (EMT) and promotes tumor progression with resultant genetic instability [45, 46]. Many of the pheno-types upon MMP3 over-expression such as genetic instability and anchorage-independent growth are shared with MEFs null for Klf4 . The up-regulation of MMPs, including MMP3 (Table 1; Figure 2), in Klf4-null cells may therefore be responsible for at least some of these events.
One of the most up-regulated genes in Klf4-null MEFs from the microarray analysis is small ubiquitin-like modifier 3 (SUMO3) (Table 1). Western blot analysis confirmed its elevation in Klf4-null cells (Figure 2). Post-translational modification by SUMOs is usually transient and alters protein function by affecting protein-protein interaction . Recent studies indicate that the SUMO cascade is involved in the mammalian DNA damage response from genotoxic stress [48, 49]. Klf4-null MEFs contain a high level of phos-phorylated histone H2AX (γH2AX), a marker for double-strand DNA breaks, and exhibit chromosome aberrations including dicentric chromosomes, double minute chromosomes, and chromatid breaks . The elevated SUMO3 levels in these cells could therefore be a reflection of the cellular response to wide-spread DNA damage observed in Klf4-null MEFs.
KLF4 is one of several factors capable of reprogramming somatic cells to induced pluripotent stem (iPS) cells [15-20]. However, the mechanism by which KLF4 achieves this task is not completely understood. KLF4 interacts with two other factors, Oct4 and Sox2, to promote reprogramming  and may form a core circuitry with other KLFs to regulate self-renewal of embryonic stem cells . KLF4 also suppresses expression of the tumor suppressor gene, p53 , which was recently shown to be a barrier to efficient reprogramming [53-57]. In this study, we demonstrated that the JAK-STAT signaling pathway is enriched in wild type MEFs relative to Klf4-null MEFs (Tables 2 and and5;5; Figure 1). This could provide another mechanism by which KLF4 may influence ES cell self-renewal. A previous study showed that promotion of mouse ES cell self-renewal and maintenance of pluripotency requires leukemia inhibitory factor (LIF)-stimulated STAT3 activation . KLF4 is a LIF-responsive gene and overexpression of KLF4 in ES cells results in a greater capacity to self-renew . A subsequent study demonstrated that KLF4 is activated by the JAK-STAT3 pathway, thus providing a mechanism by which LIF activates KLF4 . It is of interest to note that both STAT3 and phospho-STAT3 levels are elevated in wild type MEFs compared to Klf4-null MEFs (Figure 2), a finding that suggests KLF4 may activate STAT3 expression, thus providing a positive feedback loop to promote ES self-renewal. Similarly, the level of SOCS3 is higher in wild type MEFs than Klf4-null cells (Figure 2). SOCS3 is also a target of LIF signaling but promotes differentiation when over-expressed . KLF4 may therefore be positioned in a nodal point to mediate LIF-induced JAK-STAT signaling and to modulate the decision between self-renewal and differentiation.
In summary, our study established distinct transcriptional profiles of MEFs wild type and null for the Klf4 alleles. Functional clustering and pathway analysis identified a rich series of potential targets that may mediate KLF4's myriads of functions. In particular, our results further strengthened the previously established role of KLF4 in maintaining genetic stability and tumor suppression. Moreover, the results provided novel insights by which KLF4 may regulate somatic cell reprogramming. Studies are warranted to further substantiate the mechanisms by which KLF4 regulates these important processes.
This work was supported in part by grants from the National Institutes of Health (DK52230, DK64399, CA84197, and CA130308). EGH was an Emory Fellowships in Research and Science Teaching (FIRST) fellow.